The collection, curation, and analysis of data has always been as much of a social challenge as it is a technical one. As the statistical techniques and computational infrastructures of artificial intelligence and data science rapidly develop, we must not forget that it is people who ultimately make research possible – even in the most fully-automated, ‘data-driven’ computational pipeline. But how do we bring human-centered perspectives and cultural contexts to data-intensive, highly-automated algorithmic systems of knowledge production? In this talk, I discuss the role of qualitative and ethnographic methods in relation to computer, information, and data science. These holistic, reflexive, and meta-level approaches to studying data and computation in context help us better understand how to both support and practice data analytics at various scales. Focusing on the human contexts of data across the pipeline brings key insights and new collaborations to various issues, such as: the career paths of those who practice and support data science; the sustainability of open source communities that develop and maintain key software tools; the governance of complex, data-driven decision-making systems; and the interpretation of findings made from large-scale analyses of social data.